English

• The sorbicillinoids are a family of hexaketide structures that have been reported to be a characteristic class of metabolites from fungal sources [1]. Structurely, these compounds typically have a highly malleable hexacyclic ring and a sorbyl side chain, which usually can be polymerized together through Michael addition or Diels-Alder reaction catalysed by enzymes, forming complex carbon skeletons with highly oxygenated and polycyclic system [2-5]. To date, approximately 177 natural sorbicillinoids have been reported, and they exhibit a variety of significant biological activities such as anticancer, radical-scavenging, and antimicrobial activities [6-14]. Amongst them, trimeric sorbicillinoids are very rare in nature and only six compounds have been reported from the deep-sea fungi Phialocephala and Penicillium [15-18], which have attracted extensive attention due to their unique structures and remarkable bioactivities [6-8].

  In our ongoing endeavour to search for structurally unique and biologically interesting metabolites from fungal resources [19-21], we applied liquid chromatograph-mass spectrometer (LC-MS) analysis to screen endophytes from the traditional Chinese medicinal herb Coptis chinensis Franch. The LC/MS data of Trichoderma citrinoviride HT-9 revealed that this strain produced unusual sorbicillinoid metabolites, especially the minor ingredients with molecular weights approximately three times that of monomeric sorbicillinoids. Further high resolution electrospray ionization mass spectroscopy (HRESIMS)-guided dereplication analysis with subfractions found a unique molecule with a maximum ultraviolet (UV) absorption at 297 nm and a molecular ion [M + H]⁺ at m/z 729.3275 that did not match previously reported compounds, thus leading to obtain the first trimeric sorbicillinoid from terrestrial fungi, citrinsorbicillin A (1), along with two new monomeric sorbicillinoids citrinsorbicillins B (2) and C (3) (Fig. 1). 1 represented a unique carbon skeleton with two bicyclo[2.2.2]octanedione ring connected through an enolated carbon formed via two [4 + 2] cycloadditions. Moreover, 1 exhibited moderate inhibitory activity against human colon cancer HT29 cells. Herein, the details of their isolation, structural elucidation, and bioactivities, as well as plausible biogenetic pathways are reported.

  Figure 1

  

  Figure 1.  Structures of compounds 1–3.

  DownLoad: Full-Size Img PowerPoint

  Citrinsorbicillin A (1) was isolated as a yellow amorphous powder. The molecular formula was determined as C₄₂H₄₈O₁₁ on the basis of a positive HRESIMS peak at m/z 729.3275 [M + H]⁺ (calcd. for C₄₂H₄₉O₁₁, 729.3275), which required 19 degrees of unsaturation. The ¹H nuclear magnetic resonance (NMR) data (Table 1) clearly exhibited 45 protons, including six singlet methyls (δ_H 0.92, 1.03, 1.11, 1.15, 2.11, 2.26), three doublet of doublets methyls [δ_H 1.52 (dd, 6.5, 1.8 Hz), 1.61 (dd, 6.5, 1.8 Hz), 1.95 (dd, 6.8, 1.8 Hz)], six aliphatic methines [δ_H 2.46 (d, 1.8 Hz), 2.77 (dd, 10.1, 7.4 Hz), 3.21 (dd, 7.4, 1.8 Hz), 3.32 (dd, 10.0, 7.0 Hz), 3.45 (d, 1.8 Hz), 4.36 (dd, 7.0, 1.8 Hz)], eight olefinic protons (δ_H 4.86–7.33), one aromatic singlet proton (δ_H 7.68), and three distinct hydroxyl groups (δ_H 13.17, 14.61, 14.04). The ¹³C NMR spectrum (Table 1) displayed a total of 42 resonances, which could be assigned with the aid of HSQC and DEPT data to nine methyls (δ_C 7.6, 9.8, 9.9, 16.0, 17.4, 17.8, 18.9, 23.4, 24.4), six sp³ methines (δ_C 41.1, 46.3, 46.8, 46.9, 47.1, 49.1), nine sp² methines (δ_C 118.2, 127.9, 128.1, 129.9, 130.1, 130.6, 131.3, 139.4, 142.9), and 18 nonprotonated carbons (including nine olefinic carbons at δ_C 107.2, 107.3, 110.6, 112.2, 115.6, 159.5, 162.1, 169.6, 179.7; two oxygenated tertiary carbons at δ_C 75.0, 75.5; two quaternary carbons at δ_C 62.3, 62.9; and five keto groups at δ_C 197.0, 197.3, 203.2, 210.6, 211.1). Considering these NMR data and the molecular formula information with those of a known compound bisorbibutenolide isolated in this work [22], it was proposed that compound 1 was a sorbicillinoid homotrimer.

  Table 1

  

  Table 1.  ¹H (400 MHz) and ¹³C (100 MHz) NMR data for 1–3 in CDCl₃.

  DownLoad: CSV

  The planar structure of 1 was elucidated by comprehensive analysis of two-dimensional (2D) NMR data (Fig. 2). The key heteronuclear multiple bond connectivity (HMBC) correlations from H₃–37 to C-1, C-2, C-6, and C-7, from H₃–38 to C-4, C-5, and C-6, and from H-4 to C-2 and C-3, along with the ¹H–¹H correlated spectroscopy (COSY) correlations of H-4/H-8/H-7 constructed the first bicyclo[2.2.2]octanedione ring. Subsequently, the ¹H–¹H COSY correlations of H₃–17/H-16/H-15/H-7 and H₃–14/H-13/H-12/H-11/H-10, and the key HMBC correlations from OH-9 to C-3, C-9, and C-10 as well as the coupling constant of these related protons indicated that an (E)-isoallyl group was connected with the bicyclo[2.2.2]octanedione at C-7 and a (E, E)-penta-1, 3-diene group was further connected by a double bond between C-9 and C-3 to construct the fragment A. Similarly, the ¹H–¹H COSY correlations of H-24/H-25/H-26 and the key HMBC correlations from H₃–39 to C-20, C-21, C-22, and C-26, from H₃–40 to C-22, C-23, and C-24, and from H-24 to C-19 and C-20 constructed the second bicyclo[2.2.2]octanedione ring. The ¹H–¹H COSY correlations of H₃–36/H-35/H-34/H-26 as well as the large coupling constant between H-34 and H-35 (15.0 Hz) further suggested that an (E)-isoallyl group was connected with the bicyclo[2.2.2]octanedione ring at C-26 forming the fragment B. Moreover, the fragment C was determined by the HMBC correlations from H₃–41 to C-29, C-30, and C-31, from H₃–42 to C-31, C-32, and C-33, from H-29 to C-27, C-28, and C-33, from OH-33 to C-28, C-32, and C-33. Finally, the key HMBC correlations from OH-18 to C-8, C-18, and C-19, from H-8 to C-18 and C-19, and from H-24 to C-18 and C-19 suggested that the fragment A fused to fragment B via the oxygenated olefinic carbon at C-18, while the fragments B and C were connected by a single bond between C-25 and C-27 based on the HMBC correlations from H-24 and H-26 to C-27 and from H-25 to C-27 and C-28. Hence, the planar structure of 1 was established.

  Figure 2

  

  Figure 2.  Key ¹H–¹H COSY, HMBC, and NOESY correlations of 1.

  DownLoad: Full-Size Img PowerPoint

  For the relative configuration of 1, the bicyclo[2.2.2]octanedione system (A) required the spatial orientation of the bridgehead-substituted H₃–37 and H-4. The nuclear overhauser effect spectroscopy (NOESY) correlations (Fig. 2) of H₃–38/H-10, H₃–38/H₃–42, H-10/H-4, and H-4/H₃–38 suggested that these protons were proximal in space, while the correlations of H₃–37/H-15 and H-15/H-8 indicated they were co-orientated. The large coupling constant of ³ J_H-7/H-8 (7.4 Hz) further indicated the trans configuration between H-7 and H-8. Moreover, the small coupling constant (1.8 Hz) between H-4 and H-8 was similar to those of other sorbicillinoid analogues [6], thus defining their spatial position relationship. Similarly, the relative configuration of the bicyclo[2.2.2]octanedione system (B) was defined by analysis of the NOESY spectrum and coupling constant as shown in Fig. 2. Finally, the single-crystal X-ray diffraction (Fig. 3) analysis (Cu Kα) with the Flack parameter of 0.08(5) further confirmed the above deduction and determined the absolute configuration of 1 as 1R, 4S, 5S, 7R, 8R, 21R, 23S, 24S, 25R and 26R.

  Figure 3

  

  Figure 3.  X-ray crystallographic analysis of 1.

  DownLoad: Full-Size Img PowerPoint

  Citrinsorbicillin B (2) was obtained as a brown amorphous powder. The molecule formula C₁₆H₁₈O₃ was established by HRESIMS at m/z 281.1154 [M + Na]⁺ (calcd. for C₁₆H₁₉O₃, 281.1154), corresponding to eight degrees of unsaturation. The ¹H NMR data (Table 1) of 1 showed signals for two singlet methyls (δ_H 2.33, 2.35), one doublets methyl [δ_H 1.64 (d, 5.5 Hz)], three olefinic protons (δ_H 5.46–6.11), one aromatic proton δ_H 7.83 (s), and four aliphatic protons (δ_H 2.43–2.69). The ¹³C NMR and HSQC data displayed 16 carbon resonances (Table 1), accounting for three methyls (δ_C 8.2, 15.9, 17.9), two sp³ methylenes (δ_C 29.8, 34.4), four sp² methines (δ_C 109.4, 124.0, 126.9, 128.7), and seven nonprotonated carbons (110.7, 117.1, 121.8, 154.7, 156.7, 168.4, 178.5). In the 2D NMR spectra (Fig. 4), the ¹H–¹H COSY correlations of H₃–14/H-13/H-12/H₂–11/H₂–10 and the key HMBC correlations from H₂–10 to C-9 and C-8 and from H-8 to C-9 and C-10, as well as the large coupling constant between H-12 and H-13 (15.4 Hz) constructed an oxidized (E)-hepta-1, 5-diene moiety. The trisubstituted m-xylene moiety was confirmed by the key HMBC correlations from H₃–15 to C-2, C-3, and C-4, from H₃–16 to C-4, C-5, and C-6, and from H-2 to C-1 and C-6. Moreover, the HMBC correlations from H-2 to C-7 and from H-8 to C-1 and C-7 indicated that the (E)-hepta-1, 5-diene moiety at C-8 was connected to the trisubstituted m-xylene moiety at C-1 by a keto group at C-7. Considering the remaining unsaturation and the chemical shifts of C-6 (δ_C 154.7) and C-9 (δ_C 168.4), the oxo bridging between C-6 and C-9 was deduced to establish the core skeleton of chromogen. Thus, the structure of 2 was constructed.

  Figure 4

  

  Figure 4.  Key ¹H–¹H COSY and HMBC correlations of 2 and 3.

  DownLoad: Full-Size Img PowerPoint

  Citrinsorbicillin C (3) was obtained as a brown oil. Its molecular formula was deduced to be C₁₅H₂₀O₅ with six degrees of unsaturation based on the protonated HRESIMS ion at m/z 303.1206 [M + Na]⁺ (calcd. for C₁₅H₂₀O₅Na, 303.1208). The ¹H NMR data (Table 1) of 3 showed signals for two singlet methyls (δ_H 2.05, 2.19), two typical triplet methylenes (δ_H 2.91, 4.09), two aromatic protons (δ_H 6.35, 7.47), and a hydroxyl group (δ_H 12.65). The ¹³C NMR and HSQC data displayed 15 carbon resonances (Table 1), accounting for two methyls (δ_C 15.1, 21.0), five sp³ methylenes (δ_C 24.3, 25.7, 28.5, 37.7, 64.4), two sp² methines (δ_C 103.1, 132.2), and six nonprotonated carbons (113.5, 116.0, 161.1, 163.5, 171.4, 204.5). Comparison of its NMR data with those of 2 revealed they shared similar structural features. The major difference was that the pyranone ring was opened and reduced in 3, and the side chain was changed into an acetylated aliphatic chain, which were confirmed by the ¹H–¹H COSY correlations (Fig. 4) of H₂–8/H₂–9/H₂–10/H₂–11/H₂–12 and the key HMBC correlations from H₃–14 to C-13, from H₂–12 to C-13, from H₂–8 to C-1 and C-7, from H₃–15 to C-2, C-3, and C-4, from OH-6 to C-1, C-5, and C-6, from H-5 to C-4 and C-6, and from H-2 to C-1 and C-7. Thus, the structure of 3 was elucidated.

  Citrinsorbicillin A (1) represents an unprecedented carbon skeleton, whose plausible biosynthetic pathway is postulated from three molecules of sorbicillion (7) by two [4 + 2] cycloadditions (Scheme 1) [6-8]. Specifically, the known precursors 7 and sorbicillinol (8) began to be biosynthesized through a polyketide biosynthetic pathway with an oxidation process [2]. Then, the first [4 + 2] cycloaddition between the sorbyl side chain in 7 at C-2′/C-3′ and the hexacyclic rings in 8 at C-3/C-6 constructed the intermediate 9, which subsequently converted to 10 through intramolecular keto-enol isomerization. Moreover, 10 coupled 8 to construct 11 via the second [4 + 2] cycloaddition, which further performed the similar keto-enol isomerization to form the backbone 1.

  Scheme 1

  

  Scheme 1.  Plausible biosynthetic pathway of 1.

  DownLoad: Full-Size Img PowerPoint

  All of the isolates were evaluated for their cytotoxic activities against three human tumor cell lines (A549, HeLa and HT29) using methyl thiazolyl tetrazolium (MTT) method [23,24]. As a result (Table S1 in Supporting information), 1 exhibited moderate cytotoxicity against HT29 cells with half maximal inhibitory concentration (IC₅₀) value of 34.0 µmol/L, compared with the positive control drug (oxaliplatin, IC₅₀ = 45.5 µmol/L). To further explore whether 1 induced growth inhibition was associated with cell cycle regulation, PI staining analysis and cell cycle distribution were assessed by flow cytometry in HT29 cells. Compared to the control group, the percentage of G0/G1 phase cells in 1 was obviously increased at the concentrations of 17 and 34 µmol/L (Figs. 5A and B). The result indicated that 1 could induce cell cycle arrest at G0/G1 phase in a dose-dependant manner. Moreover, western blot analyses showed that the level of cyclin D1 protein significantly decreased after treatment with 1 (Figs. 5C and D). These data suggested that the cytotoxicity of 1 could be mediated by inhibiting the expression of cyclin D1 to block the cell cycle.

  Figure 5

  

  Figure 5.  The cytotoxicity analyses of 1 in HT29 cells. (A) Flow cytometry analysis of HT29 cells treated with 17 and 34 µmol/L of 1 for 48 h. (B) Percentage of cell cycle distribution of HT29 cells treated with different concentrations of 1. (C) The relative expression level of cyclin D1 in HT29 cells treated with 17 and 34 µmol/L of 1. (D) Western blot analysis of cyclin D1 protein expression of 1. The data are presented as the means ± SEM of three replicates. P < 0.05, ^***P < 0.001, compared with the control group.

  DownLoad: Full-Size Img PowerPoint

  In summary, a novel homotrimeric sorbicillinoid citrinsorbicillin A (1) and two new related monomers 2 and 3 were isolated from the Coptis chinensis endophyte Trichoderma citrinoviride HT-9. 1 possessed an unprecedented carbon skeleton and showed moderate cytotoxicity against human colon cancer HT29 cells. These findings provided a novel molecular template for chemists and pharmacologists to further design and discovery of new anticancer agents.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  Acknowledgments

  This work was supported by the National Natural Science Foundation of China (No. 82003629), the Natural Science Foundation of Chongqing (No. cstc2021jcyj-msxmX0975), and the Open Project of State Key Laboratory of Natural Medicines (No. SKLNMKF202304).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2023.109035.

  
  
• 1. [1]

     D. Lai, J. Meng, X. Zhang, et al., Org. Lett. 21 (2019) 1311–1314. doi: 10.1021/acs.orglett.8b04101

  2. [2]

     L. Kahlert, E.F. Bassiony, R.J. Cox, E.J. Skellam, Angew. Chem. Int. Ed. 59 (2020) 5816–5822. doi: 10.1002/anie.201915486

  3. [3]

     T.M. Milzarek, M. Einsiedler, H. Aldemir, et al., Org. Lett. 21 (2019) 4520–4524. doi: 10.1021/acs.orglett.9b01398

  4. [4]

     A. Sib, T.A.M. Gulder, Angew. Chem. Int. Ed. 57 (2018) 14650–14653. doi: 10.1002/anie.201802176

  5. [5]

     A. Sib, T.A.M. Gulder, Angew. Chem. Int. Ed. 56 (2017) 12888–12891. doi: 10.1002/anie.201705976

  6. [6]

     A.M. Harned, K.A. Volp, Nat. Prod. Rep. 28 (2011) 1790–1810. doi: 10.1039/c1np00039j

  7. [7]

     J. Meng, X. Wang, D. Xu, et al., Molecules 21 (2016) 715. doi: 10.3390/molecules21060715

  8. [8]

     X. Hou, X. Zhang, M. Xue, et al., J. Fungi 8 (2022) 62. doi: 10.3390/jof8010062

  9. [9]

     X.Y. Guo, H.T. Li, Y.T. Shao, et al., Fitoterapia 166 (2023) 105443. doi: 10.1016/j.fitote.2023.105443

  10. [10]

      K.A.U. Zaman, X. Wu, A.M. Sarotti, S. Cao, Nat. Prod. Res. 36 (2022) 5984–5990. doi: 10.1080/14786419.2022.2056890

  11. [11]

      X.P. Peng, G. Li, L.M. Wang, et al., Front. Microbiol. 13 (2022) 800626. doi: 10.3389/fmicb.2022.800626

  12. [12]

      X. Pang, P. Wang, S. Liao, et al., Phytochemistry 202 (2022) 113311. doi: 10.1016/j.phytochem.2022.113311

  13. [13]

      S.Z. Liu, G.X. Xu, F.M. He, et al., Mar. Drugs 20 (2022) 213. doi: 10.3390/md20030213

  14. [14]

      C. Duan, S. Wang, R. Huo, et al., J. Fungi 8 (2022) 530. doi: 10.3390/jof8050530

  15. [15]

      M.J. Cao, T. Zhu, J.T. Liu, et al., Nat. Prod. Res. 34 (2020) 2880–2886. doi: 10.1080/14786419.2019.1596099

  16. [16]

      W. Guo, J. Peng, T. Zhu, et al., J. Nat. Prod. 76 (2013) 2106–2112. doi: 10.1021/np4006647

  17. [17]

      D. Li, S. Cai, T. Zhu, et al., Tetrahedron 66 (2010) 5101–5106. doi: 10.1016/j.tet.2010.04.111

  18. [18]

      D. Li, F. Wang, X. Xiao, et al., Tetrahedron Lett. 48 (2007) 5235–5238. doi: 10.1016/j.tetlet.2007.05.134

  19. [19]

      G.P. Yin, M. Gong, G.M. Xue, et al., J. Nat. Prod. 84 (2021) 2623–2629. doi: 10.1021/acs.jnatprod.1c00185

  20. [20]

      G.P. Yin, Y.R. Wu, C. Han, et al., Org. Chem. Front. 5 (2018) 2432–2436. doi: 10.1039/c8qo00070k

  21. [21]

      G.P. Yin, Y.R. Wu, M.H. Yang, et al., Org. Lett. 19 (2017) 4058–4061. doi: 10.1021/acs.orglett.7b01823

  22. [22]

      N. Abe, T. Murata, A. Hirota, Biosci. Biotechnol. Biochem. 62 (1998) 2120–2126. doi: 10.1271/bbb.62.2120

  23. [23]

      S. Lin, J. Huang, H. Zeng, et al., Chin. Chem. Lett. 33 (2022) 4587–4594. doi: 10.1016/j.cclet.2022.03.064

  24. [24]

      Y. Peng, Y. Chang, C. Sun, et al., Chin. Chem. Lett. 33 (2022) 4261–4263. doi: 10.1016/j.cclet.2022.03.003

• 

  Figure 1  Structures of compounds 1–3.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Key ¹H–¹H COSY, HMBC, and NOESY correlations of 1.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  X-ray crystallographic analysis of 1.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Key ¹H–¹H COSY and HMBC correlations of 2 and 3.

  下载: 全尺寸图片 幻灯片
  

  Scheme 1  Plausible biosynthetic pathway of 1.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  The cytotoxicity analyses of 1 in HT29 cells. (A) Flow cytometry analysis of HT29 cells treated with 17 and 34 µmol/L of 1 for 48 h. (B) Percentage of cell cycle distribution of HT29 cells treated with different concentrations of 1. (C) The relative expression level of cyclin D1 in HT29 cells treated with 17 and 34 µmol/L of 1. (D) Western blot analysis of cyclin D1 protein expression of 1. The data are presented as the means ± SEM of three replicates. P < 0.05, ^***P < 0.001, compared with the control group.

  下载: 全尺寸图片 幻灯片

  Table 1.  ¹H (400 MHz) and ¹³C (100 MHz) NMR data for 1–3 in CDCl₃.

  下载: 导出CSV
•